Performance
Hot Spots for
AltiVec Technology
With its high-performance and ease-of-use
software environment, AltiVec technology offers
a single-chip solution to many common
embedded computing challenges. AltiVec
technology may be used to improve
performance in the following areas:
> High-bandwidth data communications
> Packet data processing
> Image and video processing
> Access concentrators/DSLAMs
ADSL and digital data concentrators
> Speech recognition
> Voice/sound processing
> Array numeric processing
> Basestation processing
> Real-time continuous speech I/O
HMM, Viterbi acceleration, neural algorithms
> 3-D graphics
Games, entertainment
High-precision CAD
> Virtual reality
> Motion video
MPEG-2, MPEG-4
H.234
> High-fidelity audio
3-D audio, AC-3, MP3
> Machine intelligence
AltiVec Technology Features
> SIMD functionality for embedded
applications with massive data
processing needs
128-bit vector execution unit with
32-entry,128-bit register file
Parallel processing with vector
permute unit and vector arithmetic
logical unit
162 additional instructions
Advanced data types, such as packed
byte, halfword and word integers, and
packed IEEE single-precision floats
Saturation arithmetic
> Simplified architecture
Virtually no interrupts other than data
storage interrupt on loads and stores
Allows hardware unaligned
access support
Virtually no penalty for running
AltiVec and standard PowerPC
instructions simultaneously
Streamlined architecture to facilitate
efficient implementation
> Maintains PowerPC ISA’s RISC
register-to-register programming model
> Supports parallel operation on byte,
halfword, word and 128-bit operands
Intra- and interelement arithmetic
instructions
Intra- and interelement
conditional instructions
Powerful permute, shift and
rotate instructions
> Vector integer and floating-point
arithmetic
Data types
8-, 16- and 32-bit signed and
unsigned integer data types
32-bit IEEE single-precision
floating-point data type
8-, 16- and 32-bit Boolean data types
(e.g., OxFFFF = 16-bit TRUE)
Modulo and saturation integer
arithmetic
32-bit “IEEE-default” single-precision
floating-point arithmetic
IEEE-default exception handling
IEEE-default “round-to-nearest”
Fast non-IEEE mode (e.g., denorms
flushed to zero)
> Control flow with highly flexible bit
manipulation engine
Compare creates field mask used by
select function
Compare RC bit enables setting
Condition Register
Trivial accept/reject in 3-D graphics
Exception detection via
software polling
Available library
*The Semiconductor Products Sector of Motorola, Inc. became Freescale Semiconductor, Inc. in 2004. Freescale and the Freescale logo are trademarks of
Freescale Semiconductor, Inc. All other product or service names are the property of their respective owners. The PowerPC name is a trademark of IBM Corp.
and used under license. Java and all other Java-based marks are trademarks or registered trademarks of Sun Microsystems, Inc. in the U.S. and other countries.
Freescale Semiconductor, Inc. 2004
ALTIVECGLANCE
REV 4
Learn More: For more information about Freescale products, please visit www.freescale.com.